Violations of expectations: Uncertainty

As cognitively efficient as heuristic processing based on interpretative frameworks may be, the world is not really as simple as it is represented in peoples’ minds (Kahneman, 2011/2012, p. 204; Lakoff, 2006, p. 13). Therefore, it is likely that people will be confronted with events or situations that do not fit their expectations. Frameworks not only serve to explain the world but also guide information selection. An attentional bias filters out information that does not match frameworks (Lakoff, 2006; Ehrlich & Maestas, 2010, p. 663; Evans, Ball, & Brooks, 1987) and anomalies will very often go unnoticed. As an example of this, Kahneman and Tversky (1982) mention the famous distorted room illusion, in which the (visual) expectation that rooms are rectangular leads observers to perceive people who are walking through the room as changing in size (p. 147). As frameworks are repeatedly challenged, however, they lose their validity. In other words, challenged self-views or worldviews suffer from instability (Van den Bos, 2009, p. 198). As I argue in the following paragraphs, such violations of expectations lead to subjective uncertainty, a negative psychological state that motivates people to reduce this through the acceptance of alternative cognitive frameworks.

Events that do not match cognitive frameworks violate the human need for coherence (Festinger, 1957/1975; Kahneman, 2011/2012). Not every such violation is necessarily negative – a surprise can also be seen as a violation of expectations (Kahneman, 2011/2012, p. 72). The main difference between surprise and uncertainty as potential outcomes of violations of expectations lies in the appreciation of the resulting situation. We may have expected a quiet night at home instead of coming home to a surprise party with

guests, music, and presents, but we are familiar with the idea of a party. Once we have accepted that the latest Peaky Blinders episode will have to wait for another night, we may enjoy this violation of expectations. For uncertainty, this is not the case. Following the violation of expectations, we find ourselves confronted with a reality that we do not know how to make sense of, and this is an uncomfortable situation. It violates our need for understanding, it leaves us uncertain about what to expect from the future, and it limits our courses of action. Uncertainty is not caused so much by the violation of expectations per se, as it is by the absence of a relevant interpretative framework.

This is different from how uncertainty is commonly defined in the literature. Knight (1921/1971) contrasts uncertainty with risk and defines it as a unique situation in which the “distribution of the outcome in a group of instances” (p. 233) is unknown. Keynes (1921/1973) acknowledges that uncertainty is characterized by unknown probabilities, but blames this on the limits to human reasoning power (p. 34), thereby essentially reducing uncertainty to risk (Blyth, 2002, pp. 31-2). Without aiming to reduce uncertainty to a merely computational problem, the notion that actors’ characteristics affect the tools with which they can react to situations does not seem completely unreasonable. In other words, both cognitive and situational aspects could affect feelings of uncertainty. As Beckert (1996) puts it:

A sociological treatment of uncertainty as a constraint on rational decision- making has to take into account both the actor and the situation. Cognitive processes are not only quantitatively limited but produce systematic biases in reasoning processes and are, moreover, socially influenced. At the same time it is also the complexity of causal relationships that leads to unintended consequences and prevents actors from optimizing behavior (p. 822).

In short, the combination of a complex situation and cognitive limits explains why actors fail to optimize their behavior.

In my definition of uncertainty, I take Beckert’s (1996) advice to focus on both actors and situations one step further. In a sense, the root of uncertainty is still to be found in both situations and actors. However, I argue that it is the interplay between actors and situations, rather than their additive effect, that gives rise to uncertainty. Just as a crisis only becomes such because of how it is perceived from a given (ideational) point of view (Blyth, 2002, p. 9; Harell, Soroka, Iyengar, & Valentino, 2012, p. 491), a situation can only be seen as complex from a given interpretative framework of reality. Uncertainty is caused by the mismatch between the expectations based on such a framework and a real situation. Therefore, we can speak of an interplay between personal and situational factors. I define the resulting feeling of not knowing how to make sense of the world as uncertainty.

The constructivist school in international relations defines uncertainty as “a lack of meaning” (Rathbun, 2007, p. 551), while Van den Bos (2009) speaks of “a subjective

sense of doubt or instability in self-views, worldviews, or the interrelation between the two” (p. 198). Both definitions are in line with the definition proposed here. My definition however has two additional advantages. First, by defining uncertainty as a result of a violation of expectations, I specify the causal mechanism behind the concept. This is very similar to the mechanism of relative deprivation, proposed by Gurr (1970/1971) to explain “why men rebel”. While Gurr’s work has been widely criticized for suffering from the ecological fallacy (Smith & Pettigrew, 2015, p. 4), his definition of relative deprivation as a result of a discrepancy between expectations (of wealth) and reality (p. 24) is a relevant analogy for the process leading to uncertainty. A second advantage of my definition is its starting point. Considering uncertainty as the violation of a more fundamental need to understand the world explains why this is such a negative state, which people wish to solve.

Institutional uncertainty

Implicit in the definition of uncertainty as the result of a violation of frameworks is the solution: adopting an alternative interpretative framework of reality. In other words, the violation of the need to understand creates a demand for new ideas to make sense of the state of the world. This mirrors the well-established theory of institutional change following uncertainty. Much like cognitive frameworks at the individual level, central to the stability of institutions and policies is the framework on which they are built:

Policymakers customarily work within a framework of ideas and standards that specifies not only the goals of policy and the kind of instruments that can be used to attain them, but also the very nature of the problems they are meant to be addressing. Like a Gestalt, this framework is embedded in the very terminology through which policymakers communicate about their work, and it is influential precisely because so much of it is taken for granted and unamenable to scrutiny as a whole. I am going to call this interpretive framework a policy paradigm (Hall, 1993, p. 279).

As we will see in what follows, such policy paradigms can of course be subject to change. However, institutions are usually not. Weyland (2008) explains this institutional stasis from a cognitive perspective, using prospect theory (Kahneman & Tversky, 1979). A key finding in experiments based on prospect theory is that “losses loom larger than gains” (Tversky & Kahneman, 1992, p. 298). As a consequence, agents who interpret a situation

in terms of losses are typically more risk-seeking, whereas the opposite is true for agents who interpret a situation in terms of gains (Weyland, 2008, p. 286-7). Hence, whether a situation is perceived in terms of gains or losses can profoundly alter people’s behavior and this is true for the institutional as much as for the individual level. Since institutional change can be highly risky (ibid.), institutions are generally not willing to change their

policy paradigms and, hence, they are characterized by stasis. This tendency is further strengthened by “[t]he distributional consequences of many institutional changes” (ibid., p. 288), which lead certain sectors to prefer the status quo and oppose change.

The fact that institutions are generally resistant to change, however, does not mean that nothing ever changes. Rather, within policy paradigms, small adjustments to solve problems are continuously made. This kind of incremental change is called first and second order change:

First and second order change can be seen as cases of “normal policymaking,” namely of a process that adjusts policy without challenging the overall terms of a given policy paradigm (...) The process of first order change is likely to display the features of incrementalism, satisficing, and routinized decision making that we normally associate with the policy process. Second order change and the development of new policy instruments may move one step beyond in the direction of strategic action (Hall, 1993, p. 279-80).

While first and second order change can serve to adjust paradigms to situations, risk aversion can prevent policy makers from making more drastic changes for a long time, even when the situation may call for it (Weyland, 2008, p. 287-9).

This can only last so long. As Hall (1993) puts it:

Like scientific paradigms, a policy paradigm can be threatened by the appearance of anomalies, namely by developments that are not fully comprehensible, even as puzzles, within the terms of the paradigm. As these accumulate, ad hoc attempts are generally made to stretch the terms of the paradigm to cover them, but this gradually undermines the intellectual coherence and precision of the original paradigm (p. 280).

In other words, if situations emerge that cannot be explained in terms of the current framework, actors will have difficulty making sense of the world through that framework and as a result it will lose its validity. A state of affairs in which paradigms are undermined can be defined as a crisis – a situation that challenges the “cognitive basis of existing institutions” (Culpepper, 2008, p. 5).

As with a discrepancy between frameworks and reality on an individual level, situations in which policy paradigms are challenged are bound to cause uncertainty. Indeed, it is deep feelings of uncertainty that typically characterize crises (Culpepper, 2008, p. 3; Blyth, 2002, p. 9). In such situations, the time is ripe for third order change; “radical changes in the overarching terms of policy discourse associated with a “paradigm shift”” (Hall, 1993, p. 279). Institutional change usually proceeds through three stages. The first stage is acknowledgment of a crisis, after which the second stage – a period of institutional experimentation (which is still characterized by uncertainty) – follows. Finally, in the

third stage, new frameworks are accepted and “actors go back to ‘taking for granted’ the game they are playing” (Culpepper, 2008, p. 5). While ideas serve to disqualify existing institutions and to define a crisis, they are also instrumental in solving crises and restoring certainty (ibid., p. 8). In situations of institutional uncertainty, ideas can guide the reinterpretation of the situation at hand and thereby they serve as “institutional blueprints” (Blyth, 2002, p. 40). As such, besides serving as frameworks, they also provide directions for future action by helping to define expectations. It follows that third order change can only take place when ideas (i.e. solutions) are available (Weyland, 2008, p. 296; Blyth, 2012, p. 208).

While Weyland (2008) sticks with the cognitive approach and describes selection of ideas as a process that is guided by heuristics (pp. 291-3), other authors focus on the appeal of ideas in terms of correspondence to “the most salient features of an old dilemma” and “crossover appeal to both negotiating partners” (Culpepper, 2008, p. 7), as well as to questions of authority (Hall, 1993, p. 280; Blyth, 2012, p. 210). Importantly, the winning idea is not necessarily the “true idea” (Culpepper, 2008, p. 3). Rather, “whether an economic idea is deemed to be “true” or not depends on how widely it is held” (Blyth, 2002, p. 33). As Hall (1993) puts it, “the process whereby one policy paradigm comes to replace another is likely to be more sociological than scientific” (p. 280), and, as a consequence, the facts are of lesser importance than authority (Blyth, 2012, pp. 210-1). In short, truth itself is socially constructed (Blyth, 2012, pp. 202-3) and as long as a framework of ideas can serve to explain the factual world and define possible courses of action, it can reduce uncertainty and reinstate institutional stability.